Which Is The Best Thermal Compoud?

There are lot of thermal compound available in ebay with fancy names e.g. silicon thermal grease, silver compound , aluminium compounds.

Each of them is claimed to be the best ..... can somebody tell me which is best for a core2duo processor..........:confused2:

thanks in advance !!!!!

 

Most will have minimal differences; the only stuff you don't want to use is the cheapie silicone-based stuff similar to what Radio Shack sells for $2.99 a tube. That stuff was fine in the old days, but tends to separate over time and become less effective.

I use two types. Arctic Silver Ceramic compound is relatively inexpensive, and can be bought in large syringe-type applicators. It works well, and has the advantage of being non-conducting, so if you make a mess with it, no biggie. I use it on all my clients' systems.

Arctic Silver 5 is what I use on my personal systems. You need to be a little more careful with it, and it's a lot more expensive. It provides very slightly better results, but remember, we're talking about 1-2 degrees C here, so if you're not overclocking your system, I'd save my shekels and get the Ceramic stuff.

The most important thing is your application of thermal compound --just enough to do the job with a very thin, even layer of it, nothing more. Use too much, and you'll raise temperatures rather than lowering them.

 

these days it does not make much of a difference which one of the artic one you have or even if yo have the stock stuff that camr with the processor

 

Incorrect. Artic Silver 5 is a far better compound for thermal transfers, but would require periodic reapplication as it will wear down after awhile, the stuff that come stock with Intel Heatsinks are made to last a lot longer but sacrifices optimal thermal transferability.



Here is a graph that compares a bunch of thermal compounds, and all of these perform better than the stock goop that comes with the Intel solution.

 

Here are more information that directly compare with stock AMD solutions which is pretty much the same as what Intel uses. Read review here



I highly recommend both Zalman STG-1 or Artic Silver 5 as I have used both with excellent results.

 

I would also avoid the Dynex stuff bestbuy has. It always turns to powder for me

 

Hi.

I like ZALMAN ZM-STG1, comes in a little glass bottle, you coat both surfaces with a brush

Very easy to apply.

cons : not cheap.

regards

John.

http://www.zalman.co.kr/ENG/product/Product_Read.asp?idx=221

 

I agree with LoneWolf onthe Ceramique, as much as it's slightly less thermally conductive, the safe factor of being electrically non-conductive is a big bonus for people not familiar with using thermal compounds. From what I've seen too it does a better job on non-lapped surfaces.

 

Just playing devils advocate here, but that graph doesnt even have the Intel compound on their comparison.

And what does that graph measure exactly? Can someone translate because I have no idea what the title says or what the 0-8 is supposed to represent.

 

No worries be the devil's advocate , I was expecting something like this but I just can't seem to find a review with stock Intel compound.

Anyway, I ended up finding a AS5 review that compares to stock compound and you should just read it through and compare to the other graph results I have provided above. As for the graphs, unfortunately the review is quite dated and the enlarged graphs no longer works but the temperatures are indirectly stated in the content of the review.



Use the lower graph for a comparison table, I only used the one at the top as it had more Thermal compounds in that review roundup despite it being a different language all together.


The idea here is to have a perspective idea on how much thermal compound can improve thermal transfers and overall cooling, as to which one is the best is up tp you to decide.

 

" Arctic Silver 5 will not separate, run, migrate, or bleed."
-actual quote from Arctic Silver's website. And since it takes 200 hours alone to cure, I don't agree with your assessment.

Your graph doesn't tell what it measures, and is in a foreign language. It could be a measure of thermal conductivity, which would tell us that yes, Arctic Silver 5 has a high measure of conductivity --however, that doesn't always translate to lower temperatures, especially over a small surface area such as a CPU die. Which means that at most, you'll get a difference of 1-2 degrees C going from something like Arctic Silver Ceramique to Arctic Silver 5 on a small die like the Merom or Penryn-core mobile Core2 processors.

Since Arctic Silver 5 also stays in a semi-liquid state, it's not one I'd recommend for laptops. Ceramique cures quicker, and to a much more solid state.

 

I don't read that language but suspect it is a measure at idle. Measure of what? I suspect 10th's of a degree if that is correct it is pretty but useless as they all perform the same for any practical purpose. Then you look at DX3 link. What does that tell us? It tells us they all work the same. The ones that are high in one category are low in the other.The problem with these tests and really the graphs is in order to make something subtle obvious they scale badly. The one that performs the worst in the first 3 performs the best in the highest stress. I know there could be explanations that explain this as being a property of the compound but I believe it is actually just variations in the sampling that can be more than explained through margin of error. Also when the graph shows a 1 degree difference it looks dramatic but in reality the true difference could be as small as 1/10 of a degree. Not the most sensitive measuring equipment. I am having trouble with any conclusions drawn from these as when I look at I see they are all the same. The pretty graphs need to be interpreted correctly. You can make a graph where the difference of one one millionth looks large. That is what is happening. Facts are not dangerous but the gathering and interpretation can be! Also not impressed (unless I missed) no control, so how do we even know if any of these do anything? I know they do but just shows very bad design. And if they do the substantial difference vs control would put the graphs in a proper (accurate) perspective. The Arctic silver document does talk of separation during storage which of course will happen as the compound is a colloid mixture but in it's application filling microscopic voids I doubt this happens. Link to Arctic Silver Document that mentions separation during storage. Chapter 5.

 

Right, I'm not using those graphs as any point of reference points and I expect a degree of error, whether it's human error or source data error. I don't take these as any means of proving anything. However, you should definitely read the links that I have provided rather than looking at pretty pictures as those may be far more meaningful than graphs alone.

I don't quite understand the cure comment you made and how it makes the Artic silver ceramique better in your terms. You can say AS5 is a liquid, but it's very thick in viscosity. Pretty much as thick as smooth peanut butter. The compound stays the same, but the heat transfer ability is optimized after it cures (does not mean dry out in any way if that's what your thinking).

1-2 degree temperature is quite significant mind you. Especially if it's at idle, that would translate to almost 5-6 degree difference on full load.

Ontop of that, it's even far more important on small die surface areas, under poor thermal conductivity the performance of a heatsink cooler would be severely hampered. Compared to desktops which have a larger heatspreader the severity is much lessoned since there's a larger surface touching the heatsink therefore better thermal conductivity. So in otherwords I think it's even more important for notebook cpu's to use decent thermal compounds for better heat management.

I see nothing wrong using Artic Silver 5 for notebooks, been using it for years for my previous dv1000, Z70VA, W3V and now my C90S. Never had issues and even if there is a higher electric conductivity rate compared Artic Silver Ceramique, CPU and GPU dies do not have any electro sensitive components so there is absolutely no risk if proper care is made to apply the compound.

 

I'am using Antec Formula 5 from Bestbuy and my CPU idles at 31c the lowest I've seen for a T9300, so I'am happy and recommend it.

 

I take it they don't sell the dynex stuff anymore?

 

Bah, just realized, you're ALL wrong !! (and so was I)

Looks like AC replaced MX-1 which was similar to Ceramique's performance, with MX-2, which destroys ACS5 while maintaining all the benefits of Ceramique/MX-1;

http://www.arctic-cooling.com/further_prod2.php?idx=140

There's your clear winner folks, according to AC themselves, good-night everybody!

 

Looks like most people don't bother reading. Check out this review here which compares the MX-2, AS5 and pretty much a slew others and the results are pretty darn close.

http://www.bjorn3d.com/read.php?cID=1133&pageID=3883&Content_Random=12088



To be honest, I don't think there is a clear cut winner and it has to do with the quality of the application rather than the compound, but you do need one of these quality products.

 

hm.. ive never changed the thermal paste on my laptop before but from my experience, MX-2 works much better then AS5 for me

But those are from desktop and i tend to change out my parts often

 

We need these tests done with a control, until then for all we know this stuff does nothing. I know it does help, I know the concept. But to present as some kind of experiment with such fundamental flaws, well just wrong. Pretending to be scientific while following none of the rules of "scientific method" bothers me. And all of you keep walking along. At this point the untreated heatsink system might operate the same as these, won't know until tested. Anecdotal information leads to incorrect conclusions.

Scale the graph to a hundred (%) all results will look the same. Hmm?

 

I totally agree with you, but there is no "good" reviews that understand the concept of a controlled test. All they tend to do is do comparison of other thermal compounds rather than comparing to a fixed rig like how overclockers.com does their heatsink comparison chart.

Until then, it's all just comparative reference.

 

Thank you, you have no idea how much better I feel that someone finally heard me. But or so. What about this.

See my point on this? The graph distorts. Like I said earlier the real diff could be 1/10 of 1 degree. Looks real large on the graph. On a % no diff.

 

Yeah exactly, those are the same questions I pose when I read through them.

If I was doing the tests, I would probably use a 100w TEC unit with a fixed heatspreader core of some sort no bigger than 1"x1", then use a giant heatsink heatpipe tower(scythe comes to mind) and fan unit 120mm, and using a thermal gun for accurate reading, so essentially taking away all possible bottlenecks from the equation. Then use controlled tests; no thermal paste, stock pink thermal gunk bubblegum looking stuff intel uses, grey thermal gunk that AMD uses, and the generic white thermal paste. Using one of the stock gunks as the "control", all the rest of the reviewed products will be based on the control number... Similar to this with resulting numbers.


Anyway, I may be able to do something like this but it would be way too time consuming.

But the bottom line is that they work, and the best way to find out is giving it a try. Most users have no idea how much thermal pastes can improve your temperatures.

 

Yes but they all seem to perform the same for any realistic purpose as the one that performs the worst at first performs the best on the OC? that 1 degree is meaningless. It could really be 1/10 of a degree diff. And by the way the AS5, the Holy Grail I think won none? So? Hmm? They all work the same, all are fine this discussion is splitting hairs based on data that broke through doors, so how can you or anyone draw an accurate informed intelligent conclusion based on this sloppy clearly inaccurate data?

 

D3X, I understand that, and actually I've seen a few that show Ceramique outperforming ACS5 (it's not just the application but also a question of the surface area IMO where conductivity and resistance play different roles), it was pretty much tongue in cheek, based on the back and forth. I personally have AS5 and Ceramique, one for CPUs the other for VPUs and chipsets.

To me someone looking for what to use should stick with the Ceramique and MX compounds simply because most people's first application is ridiculously over zealous, and when it comes to a laptop I wouldn't suggest AS5 to anyone, because even your own graph shows there's little difference in actual cooling power.

The funny thing about your graph is that it makes a 1-2 degree difference (in a graph of 1 degree steps) look like a large difference when in reality it's a margin of error difference at best. It kinda supports the idea that you can get all the cooling benefits with a non-conductive thermal interface, which wasn't always the case.

 

Correction, its not my graph

Yeah, I've done a bit of research and was pleasantly surprised with Artic Silver Ceramique performance level, it's pretty much on par with AS5.

And I totally agree with what your saying, I didn't make those graphs and I didn't do the reviews, however I do think that you should look at the numbers instead of the pretty bars and make your own judgement call.

 

LMAO, ha, ha. I know you do know what is up but it is still funny. I do agree with you, but presentation and interpretation are important! So see you around.

 

Arctic Silver MX-1 for me.

Easiest to apply and gets the best temps.

MX-2 may have replaced it, but I don't like the application method.

 

My personal favourite is Arctic Cooling MX-2, comes very close to AS5 performance wise, buts its non conductive, so if you make a mistake in application it doesn't cost you dearly. So its quite a good thermal compound for beginners

 

graph title: the difference in temperature between a cpu @ idle and a cpu@ load

i guess this is one instance where it helps being polish...

i personally use no-name silicone stuff, it works fine and it stays "fresh" for 5 years so i dont have to re-apply it once a year (like AS5 or similar, which turns into solid eventually). in terms of performace, get mx-2, for ease of use, well, i chose silicone.

 

Here are two large thermal combound roundups:
- http://benchmarkreviews.com/index.php?option=com_content&task=view&id=138&Itemid=1
- http://hardwarelogic.com/news/137/ARTICLE/2752/1/2008-03-03.html